def main(args=None):
# parse arguments
if args is None:
args = sys.argv[1:]
args = parse_args(args)
if DEBUG_MODE:
args.steps = 10
# create object that stores backbone information
backbone = models.backbone(args.backbone)
# make sure keras is the minimum required version
check_keras_version()
# optionally choose specific GPU
use_cpu = False
if args.gpu:
gpu_num = args.gpu
else:
gpu_num = str(0)
if use_cpu:
os.environ["CUDA_VISIBLE_DEVICES"] = str(666)
else:
os.environ["CUDA_VISIBLE_DEVICES"] = gpu_num
keras.backend.tensorflow_backend.set_session(get_session())
# Weights and logs saves in a new locations
stmp = time.strftime("%c").replace(" ", "_")
args.snapshot_path = os.path.join(args.snapshot_path, stmp)
args.tensorboard_dir = os.path.join(args.tensorboard_dir, stmp)
print("Weights will be saved in {}".format(args.snapshot_path))
print("Logs will be saved in {}".format(args.tensorboard_dir))
create_folder(args.snapshot_path)
create_folder(args.tensorboard_dir)
# create the generators
train_generator, validation_generator = create_generators(args)
print('train_size:{},val_size:{}'.format(train_generator.size(),
validation_generator.size()))
# create the model
if args.snapshot is not None:
print('Loading model, this may take a second...')
model = models.load_model(args.snapshot, backbone_name=args.backbone)
training_model = model
prediction_model = retinanet_bbox(model=model)
else:
weights = os.path.join(os.path.join(root_dir(), args.weights))
# default to imagenet if nothing else is specified
if weights is None and args.imagenet_weights:
weights = backbone.download_imagenet()
print('Creating model, this may take a second...')
model, training_model, prediction_model = create_models(
backbone_retinanet=backbone.retinanet,
num_classes=train_generator.num_classes(),
weights=weights,
multi_gpu=args.multi_gpu,
freeze_backbone=args.freeze_backbone)
# print model summary
# print(model.summary())
# this lets the generator compute backbone layer shapes using the actual backbone model
if 'vgg' in args.backbone or 'densenet' in args.backbone:
compute_anchor_targets = functools.partial(
anchor_targets_bbox, shapes_callback=make_shapes_callback(model))
train_generator.compute_anchor_targets = compute_anchor_targets
if validation_generator is not None:
validation_generator.compute_anchor_targets = compute_anchor_targets
# create the callbacks
callbacks = create_callbacks(
model,
training_model,
prediction_model,
validation_generator,
args,
)
# start training
training_model.fit_generator(generator=train_generator,
steps_per_epoch=args.steps,
epochs=args.epochs,
verbose=1,
callbacks=callbacks,
validation_data=validation_generator,
validation_steps=validation_generator.size())
def predict(generator,
model,
score_threshold=0.05,
max_detections=9999,
save_path=None,
hard_score_rate=1.):
all_detections = [[None for i in range(generator.num_classes())]
for j in range(generator.size())]
csv_data_lst = []
csv_data_lst.append(
['image_id', 'x1', 'y1', 'x2', 'y2', 'confidence', 'hard_score'])
result_dir = os.path.join(root_dir(), 'results')
create_folder(result_dir)
timestamp = datetime.datetime.utcnow()
res_file = result_dir + '/detections_output_iou_{}_{}.csv'.format(
hard_score_rate, timestamp)
for i in range(generator.size()):
image_name = os.path.join(
generator.image_path(i).split(os.path.sep)[-2],
generator.image_path(i).split(os.path.sep)[-1])
raw_image = generator.load_image(i)
image = generator.preprocess_image(raw_image.copy())
image, scale = generator.resize_image(image)
# run network
boxes, hard_scores, labels, soft_scores = model.predict_on_batch(
np.expand_dims(image, axis=0))
soft_scores = np.squeeze(soft_scores, axis=-1)
soft_scores = hard_score_rate * hard_scores + (
1 - hard_score_rate) * soft_scores
# correct boxes for image scale
boxes /= scale
# select indices which have a score above the threshold
indices = np.where(hard_scores[0, :] > score_threshold)[0]
# select those scores
scores = soft_scores[0][indices]
hard_scores = hard_scores[0][indices]
# find the order with which to sort the scores
scores_sort = np.argsort(-scores)[:max_detections]
# select detections
image_boxes = boxes[0, indices[scores_sort], :]
image_scores = scores[scores_sort]
image_hard_scores = hard_scores[scores_sort]
image_labels = labels[0, indices[scores_sort]]
image_detections = np.concatenate([
image_boxes,
np.expand_dims(image_scores, axis=1),
np.expand_dims(image_labels, axis=1)
],
axis=1)
results = np.concatenate([
image_boxes,
np.expand_dims(image_scores, axis=1),
np.expand_dims(image_hard_scores, axis=1),
np.expand_dims(image_labels, axis=1)
],
axis=1)
filtered_data = EmMerger.merge_detections(image_name, results)
filtered_boxes = []
filtered_scores = []
filtered_labels = []
for ind, detection in filtered_data.iterrows():
box = np.asarray([
detection['x1'], detection['y1'], detection['x2'],
detection['y2']
])
filtered_boxes.append(box)
filtered_scores.append(detection['confidence'])
filtered_labels.append('{0:.2f}'.format(detection['hard_score']))
row = [
image_name, detection['x1'], detection['y1'], detection['x2'],
detection['y2'], detection['confidence'],
detection['hard_score']
]
csv_data_lst.append(row)
if save_path is not None:
create_folder(save_path)
draw_annotations(raw_image,
generator.load_annotations(i),
label_to_name=generator.label_to_name)
draw_detections(raw_image,
np.asarray(filtered_boxes),
np.asarray(filtered_scores),
np.asarray(filtered_labels),
color=(0, 0, 255))
cv2.imwrite(os.path.join(save_path, '{}.png'.format(i)), raw_image)
# copy detections to all_detections
for label in range(generator.num_classes()):
all_detections[i][label] = image_detections[
image_detections[:, -1] == label, :-1]
print('{}/{}'.format(i + 1, generator.size()), end='\r')
# Save annotations csv file
with open(res_file, 'wb') as fl_csv:
writer = csv.writer(fl_csv)
writer.writerows(csv_data_lst)
print("Saved output.csv file")